The Student Room Group

Functions & Relations Question - Need help...

Hey guys, I was hoping someone could explain how to answer this question as I am pretty much clueless. Any help is appreciated!


Consider the relation on the set of integers Z defined by:

aRb if ab>0 or a=b=0

a. Show that this is an equivalence relation.
b. Find the equivalence classes of this relation.


This question is copied exactly from the past paper.

Thanks!

Scroll to see replies

Reply 1
Original post by cheeselike
Hey guys, I was hoping someone could explain how to answer this question as I am pretty much clueless. Any help is appreciated!


Consider the relation on the set of integers Z defined by:

aRb if ab>0 or a=b=0

a. Show that this is an equivalence relation.
b. Find the equivalence classes of this relation.


This question is copied exactly from the past paper.

Thanks!


For part a:
What three properties does a relation need to have for it to be an equivalence relation? Can you show that the relation R satisfies these three properties?

For part b:
In general, there are lots of ways of finding the equivalence classes - remember an equivalence class (in this case) is just a subset of Z such that if we pick any two elements x and y in that subset, we have xRy. One possible way of finding the equivalence classes is to pick an element of Z, and find all the elements which are in the same equivalence class as it. Now, pick another element of Z, and repeat this procedure. Keep doing this until all the elements of Z are in some equivalence class.
Reply 2
Original post by Mark13
For part a:
What three properties does a relation need to have for it to be an equivalence relation? Can you show that the relation R satisfies these three properties?

For part b:
In general, there are lots of ways of finding the equivalence classes - remember an equivalence class (in this case) is just a subset of Z such that if we pick any two elements x and y in that subset, we have xRy. One possible way of finding the equivalence classes is to pick an element of Z, and find all the elements which are in the same equivalence class as it. Now, pick another element of Z, and repeat this procedure. Keep doing this until all the elements of Z are in some equivalence class.


For part a: Does that mean i have to show that it satisfy the three axioms? If i do so, does that prove that its an equivalence relation? And i am still a little confused about part b.

Can you please write out your answer for this question?

Thank you for the help btw!
Reply 3
Original post by cheeselike
For part a: Does that mean i have to show that it satisfy the three axioms? If i do so, does that prove that its an equivalence relation? And i am still a little confused about part b.

Can you please write out your answer for this question?

Thank you for the help btw!


For part a, the definition of an equivalence relation is a relation that satisfies the three axioms you're talking about. So, if you can show that the relation does satisfy the three axioms, then it must be an equivalence relation.


First off, if we have an equivalence relation R on a set S, then an equivalence class (let's call it T) is a subset of S such that if we pick an element in T (let's call it x), then all the elements that x relates to are also in T. (i.e. if xRy, then y must be in T too). Also, if we pick two elements a and b in T, then we must have aRb.

I won't answer your question, but I'll show you how the idea of finding equivalence classes work in a different example:

Let R be a relation on Z\mathbb{Z}, defined by aRb iff a and b are both even, or if a and b are both odd. We can show that this is an equivalence relation by showing that R satisfies the three axioms.

Let's find the equivalence classes for this equivalence relation. First we'll look for the equivalence class that contains the number 2 (this is an arbitrary choice). Recall that aRb iff a and b are both even, or both odd. So, 2Rb iff b is even. Therefore, all the even integers must be in the same equivalence class as 2, and none of the odd integers can be. So one of the equivalence classes of this relation is the set of even integers. Now let's look for the equivalence class that conatins the number 1 (since we haven't found the equivalence class that this number belongs to yet). By the same logic as above, all the odd integers must be in the same equivalence class as 1, and none of the even integers are. Therefore, the equivalence classes of this relation are the set of odd integers and the set of even integers.
Reply 4
Original post by Mark13
For part a, the definition of an equivalence relation is a relation that satisfies the three axioms you're talking about. So, if you can show that the relation does satisfy the three axioms, then it must be an equivalence relation.


To show them as equivalence relations through the three axioms, would these be correct as an answer?

For ab>0

Associativity: (a) x b = a x (b) = 1 x 2 = 2 x 1 (a = 1, b = 2)

Identity: a x e = e x a = e = 1 x 1 = 1 x 1 = 1 (e = 1)

Inverse: a x a^-1 = a^-1 x a = e = 1 x 1 = 1 x 1 = 1 (a^-1 = 1)

For a = b = 0

Associativity: (a) x b = a x (b) = 0 x 0 = 0 x 0 (a = b = 0)

Identity: a x e = e x a = e = 0 x 0 = 0 x 0 = 0 (e = 0)

Inverse: a x a^-1 = a^-1 x a = e = 0 x 0 = 0 x 0 = 0 (a^-1 = 0)


I'm sorry about the late reply, because your help has been excellent! So glad that Sky isn't my broadband provider anymore!
Reply 5
Original post by cheeselike
To show them as equivalence relations through the three axioms, would these be correct as an answer?

For ab>0

Associativity: (a) x b = a x (b) = 1 x 2 = 2 x 1 (a = 1, b = 2)

Identity: a x e = e x a = e = 1 x 1 = 1 x 1 = 1 (e = 1)

Inverse: a x a^-1 = a^-1 x a = e = 1 x 1 = 1 x 1 = 1 (a^-1 = 1)

For a = b = 0

Associativity: (a) x b = a x (b) = 0 x 0 = 0 x 0 (a = b = 0)

Identity: a x e = e x a = e = 0 x 0 = 0 x 0 = 0 (e = 0)

Inverse: a x a^-1 = a^-1 x a = e = 0 x 0 = 0 x 0 = 0 (a^-1 = 0)


I'm sorry about the late reply, because your help has been excellent! So glad that Sky isn't my broadband provider anymore!


Those are three of the four group axioms -- you're working with a relation here, not a group. A relation is an equivalence relation if it is reflexive, symmetric and transitive -- these are the three things you need to check. [I'm not even really sure what you were doing with the group axioms... for example, unless a=1 or a=-1, a1a^{-1} isn't even an integer. But this doesn't matter since group theory has nothing to do with this problem anyway.]

So for your question, to do part (a) you need to check that the relation is reflexive, symmetric and transitive. Then to do part (b), you need to think of what might constitute an equivalence class. The hint is really in the >0 bit... if a>0, what is [a] (the equivalence class of a)? If a<0, what is [a]? What is [0]? Are there any more equivalence classes?
(edited 12 years ago)
Reply 6
Original post by nuodai
Those are three of the four group axioms -- you're working with a relation here, not a group. A relation is an equivalence relation if it is reflexive, symmetric and transitive -- these are the three things you need to check. [I'm not even really sure what you were doing with the group axioms... for example, unless a=1 or a=-1, a1a^{-1} isn't even an integer. But this doesn't matter since group theory has nothing to do with this problem anyway.]

So for your question, to do part (a) you need to check that the relation is reflexive, symmetric and transitive. Then to do part (b), you need to think of what might constitute an equivalence class. The hint is really in the >0 bit... if a>0, what is [a] (the equivalence class of a)? If a<0, what is [a]? What is [0]? Are there any more equivalence classes?


Ok. I think I went off track a little... Could you show me how you would show it is an equivalence relation? Or use a similar example, because I have my notes here and they are not really helping.

I'll have a little guess:

~ is a reflexive relation if it satisfies the condition:

a ~ a , for any element a in Z

a = a
a>0 = a>0 so then it is true? :/
(edited 12 years ago)
Reply 7
Original post by cheeselike
Ok. I think I went off track a little... Could you show me how you would show it is an equivalence relation? Or use a similar example, because I have my notes here and they are not really helping.

I'll have a little guess:

~ is a reflexive relation if it satisfies the condition:

a ~ a , for any element a in Z

a=b = a=b so then it is true? :/


In general, a relation R is:
- reflexive if, for all a, aRa
- symmetric if bRa whenever aRb (i.e. aRb implies bRa)
- transitive if, whenever aRb and bRc, aRc

Take for example the relation \le on Z\mathbb{Z} (i.e. aRb if aba \le b; we just write \le instead of R). This is reflexive, since for any integer a, it's definitely true that aaa \le a. It's also transitive, since if aba \le b and bcb \le c then aca \le c. But it's not symmetric, since if a<ba<b then aba \le b, but it's not true that bab \le a in this case.

In your problem, aRb if ab>0 or a=b=0. You need to show that it's reflexive, symmetric and transitive.
Original post by cheeselike
Hey guys, I was hoping someone could explain how to answer this question as I am pretty much clueless. Any help is appreciated!


Consider the relation on the set of integers Z defined by:

aRb if ab>0 or a=b=0

a. Show that this is an equivalence relation.
b. Find the equivalence classes of this relation.


This question is copied exactly from the past paper.

Thanks!


:tongue:
Reply 9
Original post by nuodai
In general, a relation R is:
- reflexive if, for all a, aRa
- symmetric if bRa whenever aRb (i.e. aRb implies bRa)
- transitive if, whenever aRb and bRc, aRc

Take for example the relation \le on Z\mathbb{Z} (i.e. aRb if aba \le b; we just write \le instead of R). This is reflexive, since for any integer a, it's definitely true that aaa \le a. It's also transitive, since if aba \le b and bcb \le c then aca \le c. But it's not symmetric, since if a<ba<b then aba \le b, but it's not true that bab \le a in this case.

In your problem, aRb if ab>0 or a=b=0. You need to show that it's reflexive, symmetric and transitive.


Ok, so seen as it says aRb if ab>0 OR a=b=0 , do I need to prove the the three axioms for both separately?

If so, then ab>0 can't be an equivalence relation can it? Because the symmetric axiom would say that if a>b then b>a which can't be true.

Also to prove that aRb if a=b=0:

Reflexive - a=a is always true therefore it is reflexive.

Symmetric - if a=b then b=a must be true, therefore it is symmetric.

Transitive - if a=b and b=c then a=c must be true, therefore it is transitive.

All three axioms are true, therefore it is an equivalence relation.

Would the above answer be correct? ^^
(edited 12 years ago)
Reply 10
Original post by cheeselike
If so, then ab>0 can't be an equivalence relation can it? Because the symmetric axiom would say that if a>b then b>a which can't be true.


No... what does a>b and b>a have to do with the statement ab>0?

Like, it is an equivalence relation; that's why you're being asked to show that it's an equivalence relation.
Reply 11
Original post by nuodai
No... what does a>b and b>a have to do with the statement ab>0?

Like, it is an equivalence relation; that's why you're being asked to show that it's an equivalence relation.


Is the part where I showed that a=b=0 is an equivalence relation at least correct?
Reply 12
Original post by cheeselike
Is the part where I showed that a=b=0 is an equivalence relation at least correct?


That bit's fine, yup. You now need to consider the case where the elements are nonzero, using the fact that a~b if ab>0. Is it true that a~a? If a~b, is it true that b~a? If a~b and b~c, is it also true that a~c? [For each of these questions, convert the "a~b" (etc) parts into mathematical inequalities using the definition of ~.]
Reply 13
Original post by nuodai
That bit's fine, yup. You now need to consider the case where the elements are nonzero, using the fact that a~b if ab>0. Is it true that a~a? If a~b, is it true that b~a? If a~b and b~c, is it also true that a~c? [For each of these questions, convert the "a~b" (etc) parts into mathematical inequalities using the definition of ~.]


Ah, ok. So would I, for example, show the reflexive relation as:

a>0 = a>0 ?

It shows that it is positive but doesn't quite look correct.

Also, would the equivalence classes be: The positive integers, negative integers and zero then? < would that be acceptable as an answer even?

Sorry to keep bombarding you with questions btw :wink:
Reply 14
Original post by cheeselike
Ah, ok. So would I, for example, show the reflexive relation as:

a>0 = a>0 ?

It shows that it is positive but doesn't quite look correct.
It's not correct. What is the = sign there for? What is the second element being multiplied on the LHS of the inequality? Etc. For reflexivity, you know that a~b means that ab>0. So what does a~a mean? Does this always hold when a is nonzero? Use a similar method for the other two.

If you do this right, you shouldn't need any = signs anywhere.

Original post by cheeselike
Also, would the equivalence classes be: The positive integers, negative integers and zero then? < would that be acceptable as an answer even?

That's correct, but you need to give justification for the claim.
Reply 15
Original post by nuodai
It's not correct. What is the = sign there for? What is the second element being multiplied on the LHS of the inequality? Etc. For reflexivity, you know that a~b means that ab>0. So what does a~a mean? Does this always hold when a is nonzero? Use a similar method for the other two.

If you do this right, you shouldn't need any = signs anywhere.


That's correct, but you need to give justification for the claim.


Okay...

So to show ab>0 as an equivalence relation:

Reflexive: aa>0 is always true if a>0

Symmetric: if ab>0 then ba>0 must be true

Transitive: if ab>0 and bc>0 then ac>0 must be true

This is the only solution I can think of, and something is still telling me that it's not quite right :frown:

And as for the equivalence classes, how can I show justification? Mathematically or worded?
Reply 16
Original post by cheeselike
Reflexive: aa>0 is always true if a>0
You need to show that this is true for all a0a \ne 0, not just a>0a>0. How else can you write aa?

Original post by cheeselike
Symmetric: if ab>0 then ba>0 must be true
Yup... but why? [It's extremely obvious, but it's worth writing it anyway.]

Original post by cheeselike
Transitive: if ab>0 and bc>0 then ac>0 must be true
Again this is true, but you need to say why (without assuming anything about the positivity or negativity of a, b or c). This one is less obvious than the previous one, so it's definitely worth saying why it must be true.

Original post by cheeselike
And as for the equivalence classes, how can I show justification? Mathematically or worded?

Well, you should show it mathematically, but it can be worded and mathematical at the same time. Given aa, the equivalence class [a][a] is the set of real numbers bb such that ab>0ab>0. For the three different cases of aa that you identified above, derive [a][a] from this definition.
Reply 17
Original post by nuodai
You need to show that this is true for all a0a \ne 0, not just a>0a>0. How else can you write aa?


aa1>0aa^-1 >0 ?

Yup... but why? [It's extremely obvious, but it's worth writing it anyway.]


Because a positive times a positive is always positive?

Again this is true, but you need to say why (without assuming anything about the positivity or negativity of a, b or c). This one is less obvious than the previous one, so it's definitely worth saying why it must be true.


If b x c can produce a positive number, that must mean that a x c must also produce a positive number seeing as a does when multiplied with b?

I don't know how else to prove it without stating that a, b and c are all positive.

Well, you should show it mathematically, but it can be worded and mathematical at the same time. Given aa, the equivalence class [a][a] is the set of real numbers bb such that ab>0ab>0. For the three different cases of aa that you identified above, derive [a][a] from this definition.


I think I know how to do this part now
Reply 18
Original post by cheeselike
aa1>0aa^-1 >0 ? Huh? I mean how else can you write aaaa? Or, to make it more obvious, how else can you write a×aa \times a? You can't assume that a>0a>0, you can only assume that a0a \ne 0 (so aa could be negative). But it's still true that aaa\sim a. You need to say why.

Original post by cheeselike
Because a positive times a positive is always positive? No; like I say, you can't assume that aa or bb is positive. But you've said that if ab>0ab>0 then ba>0ba>0. You just need to say why this is true.

Original post by cheeselike
If b x c can produce a positive number, that must mean that a x c must also produce a positive number seeing as a does when multiplied with b?

I don't know how else to prove it without stating that a, b and c are all positive. This still isn't the right reasoning. I'll give you a hint: you're assuming that ab>0ab>0 and bc>0bc>0, which means that when you multiply abab and bcbc together, this is also greater than zero. What can you deduce from this?
(edited 12 years ago)
Reply 19
Original post by nuodai
Huh? I mean how else can you write aaaa? Or, to make it more obvious, how else can you write a×aa \times a? You can't assume that a>0a>0, you can only assume that a0a \ne 0 (so aa could be negative). But it's still true that a aa~a. You need to say why.


Oh ok. I get it now lol

No; like I say, you can't assume that aa or bb is positive. But you've said that if ab>0ab>0 then ba>0ba>0. You just need to say why this is true.


Because if ab>0, then b x a will be >0. As it is the same multiplication, no matter if elements are switched.

This still isn't the right reasoning. I'll give you a hint: you're assuming that ab>0ab>0 and bc>0bc>0, which means that when you multiply abab and bcbc together, this is also greater than zero. What can you deduce from this?


Do you mean that as in a x b and b x c OR ab x bc?

I'm guessing you mean a x b and b x c, so:
That the elements multiplied together can also both be negative?

I feel as if you're feeding me the answer but I'm still too thoughtless as to understand it :/

Quick Reply

Latest